Download Advertising Empirical Generalizations: Implications for Research

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Third-person effect wikipedia , lookup

New media studies wikipedia , lookup

Propaganda model wikipedia , lookup

Economics of digitization wikipedia , lookup

Advertising campaign wikipedia , lookup

Account planning wikipedia , lookup

Sex in advertising wikipedia , lookup

Advertising management wikipedia , lookup

Advertising wikipedia , lookup

History of advertising wikipedia , lookup

Transcript
JAR49(2) 09-036
1/7
05/26/09
11:20 am
REVISED PROOF
Page:246
Advertising Empirical Generalizations:
Implications for Research and Action
YORAM (JERRY) WIND
A special conference on empirical generalizations (EGs) in advertising led to this
The Wharton School
special issue of the Journal of Advertising Research. It also generated a representative
University of Pennsylvania
[email protected]
selection of 23 EGs that give a sense of how strong our scientific knowledge is about
advertising, and where the gaps lie. While real advances in knowledge have been
BYRON SHARP
Ehrenberg-Bass Institute
for Marketing Science
University of South
achieved, the list highlights significant knowledge gaps, particularly concerning
advertising in the new fragmented interactive-media world. We surveyed advertising
thought leaders on which of these empirical laws they felt were most important and
Australia
[email protected]
marketingscience.info
most certain. Many of our empirical laws suffer from inadequate knowledge
concerning the conditions over which they do and do not generalize.
INTRODUCTION
do not apply or are modified substantially. Yet,
Having reviewed many dozens of potential em-
even if substantial unknowns exist concerning a
pirical generalizations (EGs) about advertising from
law’s generalizability, they are of enormous value
a December 2008 Wharton conference on empiri-
to management as they offer three practical benefits:
1
cal generalizations, as well as eight other EGs
presented in the MSI book on Empirical Generalizations about Marketing Impact (Hanssens, 2009)
and ten from the study of the U.K.’s IPA Effectiveness Awards (see Binet and Field, p. 130, this
issue), the question we face is, what do we do with
this wealth of information and insights?
For the practitioners among us, we hope that
the EGs offer a valuable starting point for the
design of advertising strategies. Whether you use
the EGs as principles that guide your strategy or
as hypotheses to be tested, we are all better off
having some initial knowledge of what works
and does not work and under what conditions.
Many of our advertising laws are less well
developed than we would like. We do not know a
great deal about their generalizability, so there
may be important boundary conditions where they
• As a starting point in the development of an
advertising strategy. Starting with a truly blank
piece of paper is a horrifying (but, still, rather
fanciful) prospect.
• As an initial set of tentative rules that management can follow. Or, if they feel the conditions
under which the EG holds have changed, they
can design their advertising strategy as an experiment to test this theory. These hypotheses
can encompass all the relevant advertising decisions including the message, the creative execution, the media mix, the budget, the
campaign, the frequency, etc. For example,
Wright (p. 164) shows how to employ a welldeveloped EG concerning advertising elasticity
to determine advertising budgets.
• As a benchmark, giving us some sense of how
much change to expect (e.g., in sales or sales
effects) when the world changes or when we
1
The conference was held in Philadelphia at the Wharton School,
University of Pennsylvania, and organized by the SEI Center for
Advanced Studies in Management and the Ehrenberg-Bass Institute,
University of South Australia.
246
JOURNAL OF ADVERTISING RESEARCH
June 2009
intervene with our advertising. Even partially
developed EGs can provide such grounding.
For example, the EG offered by Erik du Plessis
DOI: 10.2501/S0021849909090369
JAR49(2) 09-036
2/7
05/26/09
11:20 am
REVISED PROOF
Page:247
ADVERTISING EMPIRICAL GENERALIZATIONS
(p. 236)—and supported by the inde-
a survey on the perceived confidence/
THE PERCEIVED CONFIDENCE/VALIDITY
pendent studies of Mueller-Lust and
validity of the EGs and their importance/
AND IMPORTANCE/RELEVANCE OF THE
Bulgrin (2004) and Bronnenberg, Dubé,
relevance. Such a survey was quickly
23 EGs
and Mela (2009)—challenges the con-
conducted on 23 selected EGs. These
Figure 1 presents the classification of the
ventional hypotheses that DVR house-
EGs are only a subset of all those pre-
23 EGs on the two dimensions of con-
holds are not affected by advertising.
sented in this issue, but they were all
fidence/validity and importance/rele-
This then highlights the importance of
perceived by their authors as representa-
vance. To aid interpretation, we divided
branding in advertisements, as articu-
tive of their articles. The respondents
the 23 EGs into thirds (top, middle, and
lated in the article by Romaniuk (p. 143).
were asked to rate each description of
bottom).
an EG, or related group of EGs, using a
EGs do not restrict one’s creativity, but
rather inform it at the same time they
10-point scale on both confidence/validity
Five EGS are perceived highest on both
dimensions:
and importance/relevance.
provide inspiration and motivation that
The 23 descriptions are presented in
• value of TV EG22 (continued clout of TV);
could enhance both the creativity and im-
Table 1 grouped by four broad topics—
• value of TV EG21 (stable TV viewing
pact of advertising.
ROI, 360-Media Planning, Value of TV,
patterns in spite of social and techno-
For the academic and industry research-
and Creative Quality. While some of the
ers among us, the current inventory of
EGs could have been classified in more
EGs offers a valuable starting point for
than one category, the current grouping is
the design of our research agenda. In do-
helpful in identifying the key focus of the
• ROI EG7 (advertising’s immediate sales
ing so, we suggest focusing on three areas:
EGs. The results of our analysis of the
effect can be large and largely depends
data for the 80 respondents are presented
on creative content);
• EGs that are important, but in which
next.
logical change);
• ROI EG23 (the advertising sales response curve is convex);
• ROI EG17 (advertising elasticity is 0.1).
we have less than full confidence (i.e.,
EGs based on few studies and/or EGs
without verification from independent
researchers);
• EGs that are important and in which
we do have considerable confidence,
but are missing a number of important
conditions (product types, market segments, countries, market conditions, etc.)
that would more fully determine if the
quantitative relationship holds;
• areas that are supported by only a few
EGs (i.e., newer advertising strategies)
that we know are important, but still
underresearched.
To provide initial insights into these
three areas, the 100⫹ invited participants
at the Wharton Conference on EGs in
Advertising 2 suggested that we conduct
2
The participants included senior advertising executives
from advertising agencies, media companies, marketing/
advertising research firms, advertisers, others in the advertising industry, and academics.
Figure 1 The Perceived Confidence and Importance of the
23 EGs
June 2009
JOURNAL OF ADVERTISING RESEARCH
247
JAR49(2) 09-036
3/7
05/26/09
11:20 am
REVISED PROOF
Page:248
ADVERTISING EMPIRICAL GENERALIZATIONS
TABLE 1
The 23 EGs Included in the Survey
EG No. Empirical Generalization
................................................................................................................................................................................................................................................
ROI
................................................................................................................................................................................................................................................
4
Advertising typically has a half-life of three to four weeks. If advertising is to be sales effective in the long term, it must show
immediate sales effects in single-source data.
Kate Newstead, Jennifer Taylor, Rachel Kennedy, Byron Sharp (p. 207)
................................................................................................................................................................................................................................................
6
Based on the established EG that advertising elasticity is approximately 0.1, net profit is optimized by setting the advertising
budget to be 10% of gross profit. If the elasticity is 0.15, then the advertising budget should be 15% of gross profit, and so on.
Malcolm Wright (p. 164)
................................................................................................................................................................................................................................................
7
Brand advertising often has a pronounced short-term sales impact (as shown in single-source data). This impact decays over
time. The most dramatic influence on short-term effect is creative copy.
Leslie Wood (p. 186)
................................................................................................................................................................................................................................................
11
Even with no clicks or minimal clicks, online display advertisements generate lift in site visitation, trademark search queries, and
lift in both online and offline sales.
Gian Fulgoni, Marie Pauline Mörn (p. 134)
................................................................................................................................................................................................................................................
13
In-store digital signage featuring “newsworthy” information (e.g., new items, seasonal offers, promotions) has a markedly favorable impact on sales. This effect is stronger for hedonic (food and entertainment) products.
Raymond R. Burke (p. 180)
................................................................................................................................................................................................................................................
14
TV advertising for consumer services follow a 70:30 rule (70% of the efforts create interest, 30% create action). And, 90% of
TV advertising for consumer services dissipates within three months (versus four months for consumer goods).
Venky Shankar
................................................................................................................................................................................................................................................
17
If advertising changes by 1%, sales or market share will change by about 0.1%. (That is, advertising elasticity is 0.1.) The
advertising elasticity is higher in Europe relative to the United States, for durables relative to nondurables, in early relative to
late stages of the product life cycle, and in print over TV. The advertising elasticity is lower in models that incorporate disaggregate data, advertising carryover, quality, and promotion relative to those that do not. The advertising elasticity is lower in multiplicative models relative to other model forms, such as the additive model. The advertising elasticity is invariant over the
measure of the dependent variable or the method of estimation.
Gerard J. Tellis (p. 240)
................................................................................................................................................................................................................................................
20
There is a greater than 50% chance that the typical TV advertising campaign will lose money both short term and long term.
The risk of losing money fluctuates over the years, but has been over 50%. The average elasticity of TV advertising has fluctuated between 0.043 and 0.163 over the past 25 years.
Ye Hu, Leonard M. Lodish, Abba M. Krieger, Babak Hayati (p. 201)
................................................................................................................................................................................................................................................
23
The advertising response curve is “convex”—the greatest marginal response is from the first exposures. As the number of
cumulative exposures in a period increases, the marginal effect of the advertising drops.
Jennifer Taylor, Rachel Kennedy, Byron Sharp (p. 198)
................................................................................................................................................................................................................................................
360
Media Planning
................................................................................................................................................................................................................................................
2
A retail store layout that makes shopping quicker results in increased shopper spending.
Herb Sorensen (p. 176)
................................................................................................................................................................................................................................................
5
Approximately 20% of word of mouth (WOM) about brands refers to paid advertising in media. The level and effectiveness of
WOM are substantially increased when stimulated, encouraged, and/or supported by advertising, increasing the probability by
about 20% that a consumer will make a strong recommendation to buy or try a product.
Ed Keller, Brad Fay (p. 154)
................................................................................................................................................................................................................................................
12
If the advertisements recently recalled were on traditional media, they were more likely to have left a positive impression than if
they were on digital media. If the consumers had a previous positive impression of the brand or product advertised for advertisements recently recalled, the advertisements were more likely to have left a positive impression, regardless of the media.
Bill Moult, Walker Smith
................................................................................................................................................................................................................................................
(continued )
248
JOURNAL OF ADVERTISING RESEARCH
June 2009
JAR49(2) 09-036
4/7
05/26/09
11:20 am
REVISED PROOF
Page:249
ADVERTISING EMPIRICAL GENERALIZATIONS
TABLE 1 (cont’d)
EG No. Empirical Generalization
................................................................................................................................................................................................................................................
360
Media Planning (continued)
................................................................................................................................................................................................................................................
15
Doubling the clutter does not halve the number of advertisements recalled. Advertisements recalled in high clutter are more
likable on average.
Peter Hammer, Erica Riebe, Rachel Kennedy (p. 159)
................................................................................................................................................................................................................................................
16
Repeat viewing is 38%, and this does not alter when a program changes time. Repeat viewing is lower for comedies than
police dramas and for low rating shows, but within these program types or ratings levels repeat viewing remains at a
consistent low or high value across time changes.
Tracey Dagger, Peter Danaher
................................................................................................................................................................................................................................................
18
Where TV, radio, and magazines (and even special interest ones) claim to attract a specific audience, the target group is
typically less than half of the media’s total audience, and rival outlets often outperform them in reaching this subsegment.
Karen Nelson-Field, Erica Riebe
................................................................................................................................................................................................................................................
19
Spaced multiple exposures (distributed) produce greater learning than repeated exposures with short intervals (massed).
Longer intervals between exposures result in better learning than shorter intervals.
Alan G. Sawyer, Hayden Noel, Chris Janiszewski (p. 193)
................................................................................................................................................................................................................................................
Value
of TV
................................................................................................................................................................................................................................................
1
Over the past 15 years, TV has not declined in its effectiveness at generating sales lift and appears to be more effective than
either online or print at generating brand awareness and recognition.
Joel Rubinson (p. 220)
................................................................................................................................................................................................................................................
3
Households with DVRs are similar to non-DVR households in the basic measures of advertising effectiveness (Recall and
Recognition).
Erik du Plessis (p. 236)
................................................................................................................................................................................................................................................
21
TV still has very high reach. Declining ratings are due to fragmentation (more channels) not to reduced TV viewing levels that
are remarkably resilient to social and technological changes and to the emergence of “new media.” Average ratings halve if
the number of channels doubles. In addition, the Double Jeopardy law applies to TV channels. Bigger channels have more
viewers, and these viewers watch the channel for more hours.
Byron Sharp, Virginia Beal, Martin Collins (p. 211)
................................................................................................................................................................................................................................................
22
Despite increase in TV channels and fragmentation of audience, TV appears to retain its perceived clout among target
audiences in Asia, Europe, and North America and holds across recent years. While the influence of digital media has grown,
it has not caused a corresponding decrease in TV perceived clout.
Oscar Jamhouri, Marek Winiarz (p. 227)
................................................................................................................................................................................................................................................
Creative
Quality
................................................................................................................................................................................................................................................
8
Advertising that communicates a unique selling proposition (USP) outperforms advertising, which does not. Ideally, the USP
should be based on an important benefit; alternatively and riskier, it could be based on a feature that clearly implies a benefit.
It is effective if it is unique in the minds of consumers even though other brands could make the same claim. However, it is
especially effective if it cannot be easily matched by competitors.
J. Scott Armstrong, Patnaik Sandeep
................................................................................................................................................................................................................................................
9
The number of times a brand visually appears in a TV commercial increases the degree of correct brand association with that
commercial.
Jenni Romaniuk (p. 143)
................................................................................................................................................................................................................................................
10
Emotional response to a TV advertisement influences both branded engagement (directly) and persuasion (indirectly), and
therefore Short-Term Sales Likelihood. This pattern holds for TV advertisements across Argentina, Brazil, and Mexico, but the
magnitude of effect is different.
Jorge Alagon, Rogelio Puente
................................................................................................................................................................................................................................................
June 2009
JOURNAL OF ADVERTISING RESEARCH
249
JAR49(2) 09-036
5/7
05/26/09
11:20 am
REVISED PROOF
Page:250
ADVERTISING EMPIRICAL GENERALIZATIONS
Three EGs are perceived to be among
egories of ROI and Value of TV. Two
for the various respondent segments as
the top third in terms of importance/
EGs—Creative EG9 (the association be-
well as the outcome of a hierarchical clus-
relevance, but with perceived lower con-
tween number of times a brand appears
tering of these data embedded in the map.
fidence/validity:
in a TV advertisement and correct recall
An examination of these results sug-
of the brand) and ROI EG13 (in-store
gests significant heterogeneity among the
• Value of TV EG1 (TV more effective than
signage)—were perceived among the low-
respondents. Each industry segment has
online or print advertising in generating
est in terms of importance/relevance, but
its own “most important” EGs. Perhaps
brand awareness and recognition);
among the highest in confidence/validity.
not surprisingly academics were more likely
While informative, these surveyed per-
to view EGs that were well established in
• ROI EG4 (advertising half life of three
to four weeks);
ceptions should not serve as the only guide
the academic literature as important (EG1,
• ROI EG20 (greater than 50 percent
to the value of an EG as input to a par-
EG11, and EG23). Advertising agencies were
chance that typical TV advertising will
ticular advertising decisions. One has also
more similar in their perceived impor-
lose money in both the short and long
to note that the perception of the impor-
tance of the EGs to the academics than
term).
tant EGs varied by segment of respon-
the other advertising industry segments.
dents. Figure 2 presents the results of a
Advertisers favored EGs that focus on ROI
Not surprisingly, all of the highly
joint space analysis of the perceived im-
(EG14 and EG36) and 360 Media (EG18,
important/relevant EGs belong to the cat-
portance of the EGs and the ideal points
EG5, and EG12). Media executives perceived
Figure 2 Joint Space Map and Cluster of the 23 EGs and Six Respondent Segments
250
JOURNAL OF ADVERTISING RESEARCH
June 2009
JAR49(2) 09-036
6/7
05/26/09
11:20 am
REVISED PROOF
Page:251
ADVERTISING EMPIRICAL GENERALIZATIONS
the Value of TV (EG21) and 360 Media
The context/conditions
traditional advertising vehicles such
(EG18) as most important and were quite
Despite the importance of specifying for
as product and store designs and
similar to executives of research firms, who
each EG the conditions under which it
experiences);
also perceived ROI (EG13) and Creative
holds, most EGs do not provide this
• creative messages;
(EG10) as important. Generally, however,
needed information. In particular, partici-
• addressing the advertiser’s specific ob-
it is hard to stereotype the groups be-
pants in the survey found information
jectives (e.g., growth, launch of new
cause there was considerable heterogene-
lacking concerning the following condi-
products, crisis, and transformation);
ity within each.
tions under which EGs might or might
• addressing other portfolio decisions such
not hold:
as required timing and sequencing of
advertising initiatives.
GAPS BETWEEN THE CURRENT EGs AND
AREA FOR WHICH THE PARTICIPANTS
• competitive context;
WISH THERE WERE EGs
• different product and service categories;
When we hosted the Wharton “Empiri-
• type of media;
cal Generalizations in Advertising” in
• the message of its creative execution
late 2008, we were struck that even though
we had a large number of very diverse
presentations covering a wide variety
(both verbal and nonverbal);
Illustrative respondents’ responses include:
• “the interaction between TV brand advertising ad click through rates”;
• the countries and especially emerging
countries;
• “the synergy among different media”;
• “the focus on how to influence people
of topics, there were obvious gaps. And
• DVR households;
through total integrated communica-
even in the areas that received strong
• economic climate;
tion efforts”;
coverage (e.g., TV advertising sales ef-
• EGs that reflect the changes in con-
fects), it was very clear that more research
sumer behavior and social interaction
is needed to uncover where our existing
within advertising media, or as one
empirical laws apply and where they do
executive stated: “It is no longer tena-
• “WOM and traditional media”;
not discover EGs regarding the new ad-
ble to consider advertising and market-
• “understanding how combining media
vertising approaches that address empow-
ing as something we do to people; most
such as TV, radio, print, online, and
ered consumers and market fragmentation,
of the effect is not a direct results of
social can improve the effectiveness of
media proliferation, technological ad-
that initial interaction.”
vances, and the current global economic
recession.
The last question on our survey of ad-
• “comparison and synergy between
different
media
and
multimedia
combinations”;
a campaign”;
• How do consumers react to advertising, and how does it vary in different
contexts, moods?
• “timing and sequencing of advertising
by media/message”;
• “incorporating ‘nonadvertising’ brand
vertising thought leaders (invited partici-
building effects relating to products de-
pants at the Wharton conference) asked,
Advertising portfolios
“To guide our research agenda, reflect-
Consistent with the findings of the Future
ing on the 23 EGs and the evolving
of Advertising Project, many of the respon-
needs of advertisers, what one or two EG
dents recognized the important shift from
• “importance of continuity of effort”;
would you like to see developed.” Con-
single media effects to 360 media and to
• new media: “The store is the next big
tent analysis of the “wish list” of our
portfolios of media and message. While
frontier”; “Which online video ad ap-
respondents, augmented by some of the
much of today’s focus is on individual or
peal is most effective and under what
discussions at the conference, and some
limited number of media, one of the ma-
circumstances?”; “Mobile.”
of our own views and initial conclu-
jor needed breakthroughs in advertising
sions from the Future of Advertising Proj-
is a change in a mental model of adver-
The measurement
ect (http://seicenter.wharton.upenn.edu/
tising to a portfolio of advertising ap-
Despite the enormous progress in measur-
project_detail.aspx?keyindex⫽15&archived
proaches. Such an approach requires
ing advertising effectiveness, measure-
signs and retail experience”;
⫽0&pagebase⫽0&pageno⫽0) suggest three
• “understanding the place of online and
offline”;
ment still is one of the major challenges
major gaps between the industry needs
• integrating all media (including tradi-
facing advertising. This is especially crit-
and what we currently know about adver-
tional media, the new media such as
ical as one recognizes the expanded defi-
tising as reflected in the inventory of EGs:
social networks and mobiles and non-
nition of advertising that includes all
June 2009
JOURNAL OF ADVERTISING RESEARCH
251
JAR49(2) 09-036
7/7
05/26/09
11:20 am
REVISED PROOF
Page:252
ADVERTISING EMPIRICAL GENERALIZATIONS
touchpoints of a company/organization
executions, and undertaking a portfolio
CBS, and the Australian Research Council. He is
with its customers and other stakehold-
perspective based on the metaphor of in-
currently writing a book, Laws of Growth (see www.
ers, including their interaction among
vestment portfolio analysis.
MarketingLawsofGrowth.com).
themselves (C2C). The importance of new
measurement needs is reflected in respon-
CONCLUSIONS
dents’ quotes such as:
We hope that the rich inventory of EGs
ACKNOWLEDGMENTS
presented in this issue will lead all those
• “The obsession with recall as an indi-
concerned with advertising to:
his help with the data analysis and to Robert
cation is a real blind spot.”
• “What are the leading indications that
We are indebted to Professor Abba Krieger for
• pay more attention to the current EGs,
evaluate them carefully as to their rel-
an ad is having an effect?”
Barocci and Mark Morris for their insightful
comments and suggestions.
evance, and initiate and support the
• How to measure “trustworthy relation-
ongoing testing of EGs;
ships”?
REFERENCES
• “Better balance between short term sales
• initiate and support the development
effects and long term brand building
of new EGs in tune with the changing
effects”
business environment, empowered con-
Binet, Les, and Peter Field. “Empirical Gen-
sumers, advances in technology, mea-
eralizations about Advertising Campaign Suc-
• “Advertising response within a broader
framework”
• “The relative role of emotional versus
surement capabilities, and a research
cess.” Journal of Advertising Research 49, 2 (2009):
philosophy consistent with continuous
130–33.
adaptive experimentation.
rational responses”
• “Advertising effect without engagement”
................................................................................................
Bronnenberg, B., J.-P. Dubé, and C. Mela.
• “What is real in neuromarketing?”
YORAM (“JERRY”) WIND is The Lauder Professor, pro-
“Do DVRs Influence Consumers’ Brand Pur-
fessor of marketing, and director of SEI Center for
chases.” Working Paper, Tilburg University, 2009.
This list offers the initial building blocks
Advanced Studies in Management; and academic
for an agenda for future research on EGs
director of The Wharton Fellows Program at the Uni-
Hanssens, Mike. Empirical Generalizations about
in advertising. To fully capture, however,
versity of Pennsylvania. He currently is leading the
Marketing Impact: What We Have Learned from
the new business environment—including
“Future of Advertising” project that has overseen this
Academic Research. Cambridge, MA: Marketing
the empowered consumers and their in-
Special Issue (http://seicenter.wharton.upenn.edu/).
Science Institute, April 2009.
teraction with the advertisers and other
................................................................................................
consumers and other stakeholders—we
BYRON SHARP is a professor of marketing science
Mueller-Lust, R., and A. Bulgrin. “Digital
have to develop new mental models of
and director of the Ehrenberg-Bass Institute at the
Video Recorders: Adoption and Impact.” Ad-
advertising. These models should include
University of South Australia. His research is funded
vertising Research Foundation’s Week of Work-
an expanded definition of advertising, a
by corporations around the world including Coca-
shops, Marriot Financial Center, New York,
greater focus on creative messages and
Cola, Mars, P&G, Kraft, Turner Broadcasting,
November 2004.
252
JOURNAL OF ADVERTISING RESEARCH
June 2009